[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [K12OSN] [OT] Routing question

On Tue, 5 Nov 2002, Chris Hobbs wrote:

> big way, and it also wants bandwidth. Sitting on the black-box behind 
> the router, download speeds from a test site ranged from 80-110KB/s, 
> varying dramatically  over the course of the download. Downloading from 
> the router itself or another machine on our network not behind the 
> router resulted in downloads at a consistent 155-165KB/s, or the full 
> T-1. These tests were back-to-back, so it wasn't luck of the draw on the 
> Internet connection.
> So, the router is definitely injecting some delays.

NIC lockup comes to mind.  pool overflow in iptables also
comes to mind.  Anything in dmesg or in the ifconfig?  Does
tcpdump show ECN's when load ramps up on extended load?  Is
window size causing fragmentation efficiency and retry delay
loss? What hardware is the next hop?  Is there a speed
mismatch bouncing the uplink between 10 and 100 base -- and
upstream from that on any serial link?

-- Russ Herrold

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]